ABSTRACT
The 20th century’s increased construction activity has led to the accumulation of bridges, forming stocks, which are deteriorating with age. Aging bridges need continuous interventions either in the form of maintenance or major rehabilitation. In order to estimate future needs and optimally allocate available budgets, models to predict the future stock condition are essential.
In the present work, a series of improvements have been performed to a novel method to probabilistically estimate the future structural condition of a bridge presented by Alogdianakis et al. (2015). For this purpose, real data maintained by the Federal Highway Administration (FHWA) for USA bridges are exploited. The method presented uses the National Bridge Inventory (NBI) data (FHWA 1995) of a single year to calibrate a probabilistic model for predicting the structural condition of a bridge over time. Thus, all bridges in the data-stock processed are used, based on their ages, to represent the condition of a single bridge during its lifetime.
Curves relating bridge age with cumulative probability for each structural condition ≤ i (i = 9,8,…,4) can be assembled (code 9 represents ‘excellent’ condition, while code 4 corresponds to ‘poor’ condition. Certain time-shifts and scalings are then applied to achieve predictions for bridge ages not covered by available data. By fitting Weibull distribution functions Fi(t) to the original and shifted data (Figure 1) and specifying some criteria for deciding bridge rehabilitation, the time left for a bridge until it reaches a structural condition, that induces a need for rehabilitation, can be probabilistically evaluated. The presented method is applied to a sample of 57,056 simply supported steel bridges of various ages, which are exposed to deicing salts and elevated humidity. Weibull CDFs fitted to completed NBI data. https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781315207681/cd556cd4-4dcf-4efe-8e29-56fc67b8bfbd/content/fig331_1.tif"/>
The applicability and usefulness of the probabilistic deterioration results obtained is demonstrated with a life cycle management example. Five different attitudes of a decision maker toward risk (from risk averser to risk lover) are considered for this purpose, which influence the time-to-rehabilitation and the number of bridge upgrades within a time frame.
